PS3 Project RSX Boost: Overclock your Retail PS3 RSX Speeds (ps3 cfw only)

[LuanTeles]

@XDeltaOne

 

[fpoker102]

a

You bricked your PS3. 20xx PS3s can't remotely get anywhere near that high.

and 2nd issue is related to ps3 2503B any suggestions Bro?

 

[Hadobedo]

Boosted the voltage of my CECH-2501A with the help of RIP-Felix's awesome video tutorials and pushed my RSX to 900/950 'stable' @ 1.25V. I played GTA V for ~3hrs and didn't spot any artifacts, however when replaying the prologue mission a few hours later I experienced a random poweroff so I lowered the voltage to 1.23125V and no artifacts nor random poweroffs since. I tried 950MHz on the RSX @ 1.25V and experienced artifacts in XMB, and increasing voltage to 1.3V lead to instability/poweroff in XMB and when loading games.

Checked syscon logs after the 1.3V poweroffs/crashes and they contained new A0801002 errors. I purchased some 6.3v 470uF tantalum caps from aliexpress, now we wait... **

EDIT: purchasing 2R5TPE470M7 caps from digikey instead of aliexpress, wasn't aware of the poor quality of the caps from ali

The one benefit of not having a disc drive on this Slim is being able to route my syscon wires right thru the disc slot :P


I tried 1000MHz on the VRAM and it just led to warped textures, glitching and eventually froze in the XMB, so I reverted back to 900/950 using recovery mode and all is well. (I initially used a regular non-noBD update PUP and it got stuck at 98% updating as this Slim doesn't have a disc drive so I power cycled as it was stuck and used the proper update. The HDD got reset but that's entirely my fault in the first place aha, it was damn near empty anyways)

 

[Crkm]

Is there any way to determine the exact manufacturing date or a datacode for a Japanese PS3?

 

[Tanzu15]

a

and 2nd issue is related to ps3 2503B any suggestions Bro?

Which issue is that exactly?

@RIP-Felix @Mitsu™ @LuanTeles mother of god. 1ghz on both. I got brave and finally pushed it to 1ghz. It's perfect in xmb but in crysis it's not stable. And crashes with black flickering. So with voltage this is an easy 1ghz core. And without voltage at all 950/1000 stable all day.

 

[fpoker102]

Which issue is that exactly?

1st by mistake i entered in fsm mode cfw then by mistake i Update the cfw in fsm mode and then i format my ps3 hdd display won't appear so then i flash my ps3 with e3 flasher but nothing showing on screen Brother Tanzu any suggestions?
this issue is facing in ps3 2503B

 

[LuanTeles]

1st by mistake i entered in fsm mode cfw then by mistake i Update the cfw in fsm mode and then i format my ps3 hdd display won't appear so then i flash my ps3 with e3 flasher but nothing showing on screen Brother Tanzu any suggestions?
this issue is facing in ps3 2503B

Sorry but it is unrecoverable :/

 

[RIP-Felix]

Boosted the voltage of my CECH-2501A with the help of RIP-Felix's awesome video tutorials and pushed my RSX to 900/950 'stable' @ 1.25V. I played GTA V for ~3hrs and didn't spot any artifacts, however when replaying the prologue mission a few hours later I experienced a random poweroff so I lowered the voltage to 1.23125V and no artifacts nor random poweroffs since. I tried 950MHz on the RSX @ 1.25V and experienced artifacts in XMB, and increasing voltage to 1.3V lead to instability/poweroff in XMB and when loading games.

Checked syscon logs after the 1.3V poweroffs/crashes and they contained new A0801002 errors. I purchased some 6.3v 470uF tantalum caps from aliexpress, now we wait...

The one benefit of not having a disc drive on this Slim is being able to route my syscon wires right thru the disc slot :P
View attachment 43695 View attachment 43696

I tried 1000MHz on the VRAM and it just led to warped textures, glitching and eventually froze in the XMB, so I reverted back to 900/950 using recovery mode and all is well. (I initially used a regular non-noBD update PUP and it got stuck at 98% updating as this Slim doesn't have a disc drive so I power cycled as it was stuck and used the proper update. The HDD got reset but that's entirely my fault in the first place aha, it was damn near empty anyways)

Don't use 6.3v caps or buy from Ali. Use these or an AlPol cap with lower ESR and 2-2.5v rating. Buy from a reputable company like LCSC, mouser digikey, arrow, farnell...etc. Otherwise, you risk getting reject bin caps with unreliable ratings. Capacitance all over the place, ESR that's way too high, endurace rating that doesnt meet 2000hr specification, leakage current, etc.

 

[fpoker102]

Sorry but it is unrecoverable :/

Ok Brother Luan I Understand

 

[ryeno]

@RedOverlord tried that with an A00 frankie that has a CXD5302. It seemed to be like most other TSMC 40s in terms of core performance.

rsx40 a01 750/1000 vpe:ff shd:3f  [PCV367-01:0:4:1:10:6:6:0:2][23:0:e:0:1:0:1][1:1:0]

But I have noticed that it differs by manufacturer.
View attachment 43624
This graph will get better with more reports of people highest stable overclock (without voltage mod preferably). But As always I need more LV1 RSX info strings to extract the data needed. I'd show the 65nm and 90nm data, but there haven't been enough of them overclocked to their max to make a graph like this. I need more data before I can say what their average is, based on fab. Even with 160 consoles worth of data. Of which most haven't been overclocked or didn't dump the LV1 so I could know who made their RSX. So this only includes those who did. Which as of yet is a relitively small number of people.

Because of this, you need to take these graphs with a huge grain of salt. However, the general trend that SONY fabbed 40s overclock higher is significant and I think reliable. As for what you can overclock them to, that's what you shouldn't take faith. Don't look at that graph and thing your sony fabbed 40nm can do 900MHs and jump strait to it. Or you'll probably brick.

did you ever release the voltage mod?

 

[Hadobedo]

Don't use 6.3v caps or buy from Ali. Use these or an AlPol cap with lower ESR and 2-2.5v rating. Buy from a reputable company like LCSC, mouser digikey, arrow, farnell...etc. Otherwise, you risk getting reject bin caps with unreliable ratings. Capacitance all over the place, ESR that's way too high, endurace rating that doesnt meet 2000hr specification, leakage current, etc.

Don't know how I missed this video, super eyeopening wow... Going to purchase some 2R5TPE470M7** caps from Digikey, thanks again for steering me in the right direction!!

EDIT: getting EEF-HX0D561R4 instead as they are cheaper

 

[ryeno]

Don't know how I missed this video, super eyeopening wow... Going to purchase some 2R5TPE470M7 caps from Digikey, thanks again for steering me in the right direction!!

Too expensive. I would buy the EEF-HX0D561R4 or EEF-SX0D331XR or similar.

 

[Hadobedo]

Too expensive. I would buy the EEF-HX0D561R4 or EEF-SX0D331XR or similar.

Woah almost a dollar cheaper per unit, thanks !

EDIT:
ultimately ended up buying 32 PANASONIC 2R5TPE470M7 from LCSC which is what RIP-Felix linked in his first reply thanks for your patience :P total was $26 USD with shipping

 

[RIP-Felix]

Don't know how I missed this video, super eyeopening wow... Going to purchase some 2R5TPE470M7** caps from Digikey, thanks again for steering me in the right direction!!

EDIT: getting EEF-HX0D561R4 instead as they are cheaper

560uF is too much.

LCSC like I linked are on sale RN, 2 for 1$. Would be like $15 + shipping to replace all the tokins on a COK (24 total).

did you ever release the voltage mod?

This is 1 of my 2 consoles. 2501A.
rsx40 a01 700/900 vpe:ff shd:3f [NN6051-11:0:4:b:15:4:6:0:1][25:0:e:0:1:0:1][1:1:0]

Yes

Spoiler: Voltage Mod Tutorial

 

[ryeno]

560uF is too much.

LCSC like I linked are on sale RN, 2 for 1$. Would be like $15 + shipping to replace all the tokins on a COK (24 total).



Yes

Spoiler: Voltage Mod Tutorial

I was going to add 3 560 onto an existing 2501 motherboard (on the backside?) to increase the capacitance per your analysis.

 

[RIP-Felix]

I was going to add 3 560 onto an existing 2501 motherboard (on the backside?) to increase the capacitance per your analysis.

So what you saw me add were the same 470uF aluminum polymer caps sony used. I forget the exact model number, but they were 470uF in the Panasonic EEF series. Probably 4.5 or 3mOhms ESR each. Technically what I did was a band-aid and the proper solution would be to remove the old caps and replace them with new ones. Preferably Tantalum for the better ESR performance, but it was effective in the short term, which is what I was looking for to evaluate the overclocks and voltage mods.

That console I purchased to harvesxt the RSX from and didn't expect it to be alive. So I wasn't really worried about proper fixes.

If you want to do the same, I would reccomend the same exact cap as wjat sony has on there. Either 470uF AlPol or 470uF TaPol.

 

[Mitsu™]

Appreciable improvement here !

 

[ryeno]

Seems like 1.1v was the stock voltage for 7800GS
1.138v/1.150v was the factory limit for the 40nm Nvidia GTX 480.
I'll probably start with 1.1v and see how the core responds and then lower voltage until i find a stable OC.

So what you saw me add were the same 470uF aluminum polymer caps sony used. I forget the exact model number, but they were 470uF in the Panasonic EEF series. Probably 4.5 or 3mOhms ESR each. Technically what I did was a band-aid and the proper solution would be to remove the old caps and replace them with new ones. Preferably Tantalum for the better ESR performance, but it was effective in the short term, which is what I was looking for to evaluate the overclocks and voltage mods.

That console I purchased to harvesxt the RSX from and didn't expect it to be alive. So I wasn't really worried about proper fixes.

If you want to do the same, I would reccomend the same exact cap as wjat sony has on there. Either 470uF AlPol or 470uF TaPol.

No, I don't think so. I think the issue is the lack of capacitance, not degraded caps. I think we need more capacitance on the 2501s because the total capacitance is lower than the NEC TOKIN 2000 series. 470uF x 7 = ~3300uF. TOKIN is 1000/1200 * 4 = 4000 or 4800uF. Adding 3 470uF brings the total capacitance back to PS3 FAT spec. I'll probably buy the EEF-SX0D471XE on Mouser and then add 3 or 4 caps.

Hello, I was able to overclock my Slim 2501A (July 2010 mfr. date) to 750MHz Core and 900MHz VRAM stable* (*so far!)

rsx40 a01 750/900 vpe:ff shd:3f [NN2316-03:0:4:3:16:5:6:0:2][24:0:a:0:1:0:1][1:1:0]

I recently delid this Slim and applied liquid metal between the IHS and CPU/RSX, with cobra dynamic fan set to 65°C and min fan speed set to 35% the CPU sits at 60-61°C and RSX sits at 62-63°C

800MHz on Core causes artifacts and I didn't try any VRAM values higher than 900MHz as it seems like there isn't much benefit/large risks when pushing VRAM further than that anyways

I didn't do much before/after benchmarks but it definitely seems like the FPS in games like Metal Gear Rising Revengeance and GT6 dips much less and less often! I'm very pleased with these results

EDIT: regarding delid: came across RIP-Felix's delid guide and now know that best practice is to use high quality thermal paste instead of liquid metal and to reglue the IHSes to maximize lifespan of the chips. will be doing that asap!!

EDIT 2: increased VRAM clocks to 950MHz and everything still seems stable, gonna leave it at that, for real this time :P

MY PS3 was game stable at VRAM 950 but sometimes it wouldn't boot games at 950. So if 900 is stable then I would recommend trying 925.

I don't think you should use liquid metal. The IHS must be engineered to resist liquid metal. It might not be safe long term.

 

[RIP-Felix]

I think we need more capacitance on the 2501s because the total capacitance is lower than the NEC TOKIN 2000 series. 470uF x 7 = ~3300uF. TOKIN is 1000/1200 * 4 = 4000 or 4800uF. Adding 3 470uF brings the total capacitance back to PS3 FAT spec. I'll probably buy the EEF-SX0D471XE on Mouser and then add 3 or 4 caps.

You have no idea how many times I've heard this excuse as copium. Not trying to be antagonistic here, but a capacitor is not a battery and capacitance is not a thing you just need to refil like a baloon that's lost too much helium to float.

I'll give you an example. Several people have seen their 1002/1001 errors return about a year after adding parasites. I have seen noise and ripple get worse when you exceed 10,000uF. I've seen adding caps result in worse filtering. I've seen removing 1 bad cap return the onsole to working order without replacing it!

A filter has to consider the frequency response curve of the total combined frequency response of the caps in the array. And if any one of them is bad it can throw the entire filter off.

I suspect this is what happened with mine when I added those parasites. Even though it looked ok, it may have thrown off the frequency reaponse enough to let certain frequencies through and one of them tripped voltage protection.

I'm still resear hing why it may have happened, but I think a good filter can easily deliver more than the 1.2v I increased it to. I'm not convinced it was a current limit.

 

[ryeno]

You have no idea how many times I've heard this excuse as copium. Not trying to be antagonistic here, but a capacitor is not a battery and capacitance is not a thing you just need to refil like a baloon that's lost too much helium to float.

I'll give you an example. Several people have seen their 1002/1001 errors return about a year after adding parasites. I have seen noise and ripple get worse when you exceed 10,000uF. I've seen adding caps result in worse filtering. I've seen removing 1 bad cap return the onsole to working order without replacing it!

A filter has to consider the frequency response curve of the total combined frequency response of the caps in the array. And if any one of them is bad it can throw the entire filter off.

I suspect this is what happened with mine when I added those parasites. Even though it looked ok, it may have thrown off the frequency reaponse enough to let certain frequencies through and one of them tripped voltage protection.

I'm still resear hing why it may have happened, but I think a good filter can easily deliver more than the 1.2v I increased it to. I'm not convinced it was a current limit.

Adding capacitance is what reduces low frequency ripple. A degraded cap would be a capacitor with low capacitance and lower capacitance means more low frequency ripple. So if you want to reduce the low frequency ripple then you must increase the total capacitance. In the 2500 series, I don't think the high ripple is because of degraded caps reducing capacitance but because the overall capacitance is low from the factory.

[

The PS3 vram is pretty budget. It's only 2 phase? So I think we'll hit the current limit very fast or start burning up our VRMs if we increase the current usage. I wouldn't raise the voltage beyond 1.1v for a daily driver, staying below 1.05v is even better.